This week the focus on algorithms turned more specifically to learning analytics (LA), with a video lecture by Ben Williamson (2014), and George Siemens’ (2013) paper on the emergence of LA as a discipline. In my own explorations, the focus has been more on how analytics might be used to improve learning, and obstacles to this, including the potential subversion of analytics by competing organisations. For the opportunities to improve learning, I wrote in response to Lockyer, Heathcote and Dawson‘s proposal for a check-points and processes framework to evaluate learning design, Durall & Gros‘ (2014) suggestion that LA could be used by students as a metacognitive tool to underpin self-directed and self-regulated learning and Wintrup‘s assertion that LA could be used by teachers and students to work on the process of learning (2017). Wintrup’s paper also contributed significantly to the development of my understanding of the risks of using LA to “improve” student engagement and learning, alerting me to a lack of alignment between the way ‘student engagement’ is used within LA and the way it is used in established research (Koh, 2001, for example). Wintrup also writes about how our ability to measure specific data points can result in these points being viewed as significant within assumptions about quality, and resultantly shaping learning in potentially unwelcome ways. This notion of analytics producing worlds rather than just reporting on them (Knox, 2015; Kitchin & Dodge, 2011) was also taken up in discussion of an article on the potential for AI to be used by authoritarian regimes, and in reference to IBM’s visions of cradle to career tracking.
The impact of how – and by whom – analytics are applied was a recurring theme within most posts this week. It also appeared in the Tweetorial – to which many of my lifestream posts this week relate. These will be unpacked more fully in week 10, but the main concerns expressed within discussions I took part in related to ownership of data, the intrusion of corporate motives into learning, value-laden assumptions implicit in LA, and the partial nature of the picture that LA offers about learning and student engagement.
Good to see you focusing on questions around learning analytics and the application of ‘data science’ education – your post on the ‘Informing Pedagogical Action’ paper was great, and I left some other comments there.
I think you make a good and useful distinction there between learning analytics in a broad sense of tracking student progress, and specific projects that are directed at teaching – the latter tend to be a little more interesting I think. However, you’re right to get critical here though, and the Wintrup paper looks really good – wasn’t aware of this, so thanks for sharing! The drive to create efficiency seems habitual in most of these projects, which is perhaps telling, predetermining the very ‘problem’ analytics is supposed to ‘solve’. Great summary here from you as well about the critical aspects of analytics.